Notebookcheck Logo

Civilization: Beyond Earth Benchmarked

Hexes live on. Sid Meier's successful gaming series moves on to the next round with Beyond Earth. We don't just mean content-wise -- in terms of graphics, too, the game is on the move. Thanks to Mantle support, AMD-based computers should have a noticeable advantage over Nvidia systems. We ran the tests and have laid out the results for you here.
Civilization: Beyond Earth Logo

For the original German article, see here.

Graphics

Like its predecessor, Beyond Earth doesn't exactly offer stunning visuals. Whether we're talking about texture quality, polygon count, or effects: The turn-based strategy game's graphics are mostly limited to the necessities. Even so, the new Civilization game gains our respect for its nice details and overall look that just works. In other disciplines, too, there's not much to critique: At a size of about 6 MB, the game proves to be relatively compact. Some of this year's games need seven times the space (The Evil WithinShadow of Mordor, Wolfenstein: The New Order).

We're fans of the video menu -- it looks clean and has a ton of detail options. Under "Graphics Profile", the overall quality can be changed with just one click. There are also switches for anti-aliasing (2-8x MSAA) and the resolution. Vertical synchronization eliminates image tearing.

Low Settings
Low Settings
Medium Settings
Medium Settings
High Settings
High Settings
Ultra Settings
Ultra Settings

Frustrating, on the other hand, is the fact that many settings changes require the game to be restarted. Not only that, but the intro videos can't be interrupted and Beyond Earth takes a long time to load the main menu (even with an SSD). Now and again, Civilization doesn't start correctly in full screen mode, though there is a trick to solve that problem (right click on the taskbar symbol "close window"). In general, the programming behind the graphics seems a little underdeveloped. The game even crashed on several systems.

One of Beyond Earth's biggest highlights is that it supports AMD's Mantle API, which competes with Microsoft's DirectX interface. Thanks to many optimizations, Mantle is supposed to work more efficiently. We did notice a small increase in performance in our tests (+5 % @ Radeon R9 M290X & R9 280X), but the improvement isn't really groundbreaking. In most cases, Nvidia can keep up quite well.

Low Settings
Low Settings
Medium Settings
Medium Settings
High Settings
High Settings
Ultra Settings
Ultra Settings

Apart from that, the game's performance is very closely tied to the GPU's raw power. Taking a look at the benchmark table, it's clear that the processor is only a limiting factor at the lowest quality setting -- at least when it comes to Intel notebooks -- and that only in systems with high-end GPUs. While the frame rate sometimes varies greatly from preset to preset (High/Ultra being exceptions), the differences in visuals are more limited. Excellent: According to our benchmarks, MSSA doesn't cost a lot in terms of performance.

The drivers we used are, of course, the current versions from Nvidia (344.48 WHQL) and AMD (14.9.2 Beta). Windows 7 users should be careful when downloading the Catalyst driver. Even though we selected the right option, the standard site ("AMD Catalyst Packages") only has a Windows 8 .exe file that doesn't function properly with Windows 7. The appropriate download is only available on the detail page.

Benchmark

Hidden benchmarks seem to be all the rage lately. Like Alien: Isolation, Beyond Earth also offers a performance test that can only be activated with a little finesse. To get to the benchmark, under "Set Launch Options" (right click on the game in Steam, select "Properties") you have to add the command -benchmark lategameview. The next time the program boots, the test will initialize automatically.

Because the benchmark provided is a worst-case-scenario segment with tons of things going on, even powerful GPUs break into a sweat (especially when it comes to the minimum fps). We were pleased with the fact that the benchmark delivered constant results -- unsurprisingly, since the sequence is 100% identical every time.

Unlocking the benchmark
Unlocking the benchmark
Opening the results
Opening the results
Converting the data
Converting the data

We were not so pleased with the fact that the game doesn't display the results at the end of the sequence. Instead of presenting the minimum, average and maximum frame rates right after the test, Beyond Earth just stores them in the program folder (SteamSteamAppscommonSid Meier's...) in a file named lategameview. At first glance, the file appears to be full of arbitrary numbers -- only after some spreadsheet analysis did we finally get our sought-after results.

Because it was simple, we used an online tool we programmed ourselves to calculate the frame rates (see above screenshot), though this tool was only intended for internal use. If necessary, you could also record the benchmark using Fraps. Except for the minimum frame rates, the results are almost identical. Unfortunately, Fraps only works in DirectX mode, not in Mantle mode.

Results

Even though Beyond Earth is no graphic wonder, the game requires a fair amount of computational power. For the lowest setting (1024 x 768), you'll need at least a GeForce GT630M, Radeon HD 8650G or HD Graphics 4600. Older entry-level GPUs like the Radeon HD 7660G or the HD Graphics 3000/4000 are already out of their league here.

Medium details and 1366 x 768 pixels can only be rendered more or less fluidly by a GeForce GT 750M or above. For the "High" preset and 1920 x 1080 pixels, we suggest a GeForce GTX 765M or better. If you want to activate the "Ultra" preset with 4x anti-aliasing, you'll need a card on par with the GTX 850M, GTX 860M or GTX 770M.

Civilization: Beyond Earth
    1920x1080 Ultra Preset AA:4x MS     1920x1080 High Preset     1366x768 Medium Preset     1024x768 Minimum Preset
GeForce GTX 980, 3770K
Desktop-PC
110.8 (52min) fps ∼78%
119.8 (54min) fps ∼77%
149 (55min) fps ∼54%
150.9 (58min) fps ∼56%
Radeon R9 280X, 3770K
Desktop-PC
65.5 (40min) fps ∼46%
72.2 (45min) fps ∼47%
103.1 (54min) fps ∼37%
160.3 (68min) fps ∼59%
GeForce GTX 880M, 4700MQ
Schenker W504
53.6 (24min) fps ∼38%
60.3 (25min) fps ∼39%
94.3 (35min) fps ∼34%
143.3 (58min) fps ∼53%
GeForce GTX 780M, 4700MQ
Schenker W503
47.6 (20min) fps ∼34%
51.4 (26min) fps ∼33%
81.9 (27min) fps ∼30%
135.9 (37min) fps ∼50%
GeForce GTX 870M, 4700MQ
Schenker W504
45.1 (19min) fps ∼32%
51.4 (22min) fps ∼33%
81.1 (30min) fps ∼29%
142.5 (57min) fps ∼53%
Radeon R9 M290X, 4700MQ
Schenker W504
44 (28min) fps ∼31%
48.9 (33min) fps ∼32%
76.5 (43min) fps ∼28%
152.7 (71min) fps ∼56%
GeForce GTX 860M, 4700MQ
Schenker W504
34.2 (15min) fps ∼24%
41.3 (18min) fps ∼27%
63.7 (24min) fps ∼23%
141.8 (52min) fps ∼52%
GeForce GTX 770M, 4700MQ
Schenker W503
30 (14min) fps ∼21%
32.7 (15min) fps ∼21%
53.4 (17min) fps ∼19%
123.6 (30min) fps ∼46%
GeForce GTX 850M, 4340M
Schenker M504
29.7 (10min) fps ∼21%
35.5 (20min) fps ∼23%
54.9 (28min) fps ∼20%
109.2 (50min) fps ∼40%
GeForce GTX 765M, 4700MQ
Schenker W503
24.7 (10min) fps ∼17%
27.8 (12min) fps ∼18%
41.7 (17min) fps ∼15%
109.8 (33min) fps ∼41%
GeForce GT 750M, 4702MQ
Schenker M503
14.7 (8min) fps ∼10%
16.8 (9min) fps ∼11%
26.4 (12min) fps ∼10%
66.7 (24min) fps ∼25%
Radeon R7 512 Cores (Kaveri Desktop), A10-7850K, Samsung SSD 470 Series MZ-5PA256/EU
A10-7850K Asus A88-XM-PLUS
14 (9min) fps ∼10%
15.6 (10min) fps ∼10%
24 (14min) fps ∼9%
65 (32min) fps ∼24%
Iris Pro Graphics 5200, 4750HQ, Intel SSD 525 Series SSDMCEAC180B3
SCHENKER S413
13.6 (7min) fps ∼10%
16.3 (9min) fps ∼11%
24.2 (10min) fps ∼9%
53.3 (12min) fps ∼20%
GeForce GT 740M, 4200M
HP Envy 15-j011sg
12 (5min) fps ∼8%
14.5 (6min) fps ∼9%
23 (11min) fps ∼8%
61.4 (29min) fps ∼23%
GeForce GT 640M, 2637M, Lite-On LMT-256M3M
Acer Aspire M3-581TG
17.4 (9min) fps ∼6%
43.8 (17min) fps ∼16%
GeForce GT 720M, 4200M, WDC Scorpio Blue WD10JPVX-22JC3T0
MSI CX61-i572M
9 (0min) fps ∼6%
15.4 (0min) fps ∼6%
44.8 (0min) fps ∼17%
HD Graphics 4600, 4700MQ
Schenker W503
7.4 (5min) fps ∼5%
8.9 (5min) fps ∼6%
13.7 (8min) fps ∼5%
37 (18min) fps ∼14%
Radeon HD 8650G, A10-5750M, Samsung SSD 470 Series MZ-5PA256/EU
Pumori Test Platform (A10-5750M)
14.9 (0min) fps ∼5%
38.7 (0min) fps ∼14%
GeForce GT 630M, 3720QM, Seagate Momentus 7200.5 ST9750420AS
Asus N56VM
6.7 (2min) fps ∼5%
7.7 (3min) fps ∼5%
13 (3min) fps ∼5%
35.6 (4min) fps ∼13%
HD Graphics 4000, 3720QM
Asus N56VM
8.1 (2min) fps ∼3%
18.2 (4min) fps ∼7%

Test Systems

Four of our test systems are courtesy of Schenker Technologies (mysn.de):

  • W504 (Core i7-4700MQ, 8 GB DDR3, GeForce GTX 860M, GTX 870M, GTX 880M, Radeon R9 M290X)
  • W503 (Core i7-4700MQ, 8 GB DDR3, GeForce GTX 765M, GTX 770M, GTX 780M)
  • M504 (Core i5-4340M, 8 GB DDR3, GeForce GTX 850M)
  • M503 (Core i7-4702MQ, 8 GB DDR3, GeForce GT 750M)

Windows 7 (64 bit) is installed on each of these notebooks.

GPU drivers used: Nvidia 344.48, AMD 14.9.2 Beta, Intel 10.18.10.3960

Further test devices are courtesy of Nvidia:

Intel also furnished us with one:

  • Schenker S413 (Core i7-4750HQ, 8 GB DDR3, Iris Pro Graphics 5200)

The desktop computers are equipped with CPUs/APUs from Intel and AMD, SSDs from Micron, Intel and Samsung, mainboards from Intel and Asus, as well as graphics cards from Nvidia and AMD.

Overview

Show Restrictions
PosModel< PrevNext >Civilization: Beyond Earth
 Civilization: Beyond Earth (2014)
low
1024x768
Minimum Preset
med.
1366x768
Medium Preset
high
1920x1080
High Preset
ultra
1920x1080
Ultra Preset
4x MSAA
NVIDIA GeForce GTX 980M SLI
175.2
172.5
154.6
141.9
NVIDIA GeForce GTX 980
150.9
149
117.3n2
108.6n2
NVIDIA GeForce GTX 970M SLI
159.4
150.9
115.2
104.6
AMD Radeon R9 290X
199
171
117
106
NVIDIA GeForce GTX 980M
132n3
105.6n3
88.1n3
76n3
AMD Radeon R9 280X
160.3
103.1
72.2
65.5
NVIDIA GeForce GTX 960
167.8
131
90.9
77.1
NVIDIA GeForce GTX 970M
158
120.4
79.75n2
69.75n2
AMD Radeon R7 370
186.4
83.5
57.2
50.4
NVIDIA GeForce GTX 880M
143.3
94.3
60.3
53.6
NVIDIA GeForce GTX 780M
135.9
81.9
51.4
47.6
NVIDIA GeForce GTX 965M
121.8
88.2
61.1
49.9
NVIDIA GeForce GTX 870M
142.5
81.1
51.4
45.1
AMD Radeon R9 M290X
152.7
76.5
48.9
44
NVIDIA GeForce GTX 960M
104.45n2
67.15n2
45.35n2
39.55n2
NVIDIA GeForce GTX 770M
123.6
53.4
32.7
30
NVIDIA GeForce GTX 860M
104.3n2
57.45n2
38.25n2
32.8n2
NVIDIA GeForce GTX 950M
55.3
37.8
30.7
NVIDIA GeForce GTX 850M
109.2n3
54.9n3
35.5n3
29.7n3
PosModel< PrevNext >Civilization: Beyond Earth
low med. high ultra
AMD Radeon R9 M280X
72.3
48.6
31.3
19
NVIDIA GeForce GTX 765M
109.8
41.7
27.8
24.7
AMD Radeon R9 M370X
85.7
39.2
26.8
22.7
AMD Radeon R9 M265X
65.9
29.8
19.9
16.8
NVIDIA GeForce GT 750M
66.7
26.4
16.8
14.7
NVIDIA GeForce 940MX
67.2
29.7
21.4
17.3
NVIDIA GeForce 940M
64.3n3
28.2n3
18.9n3
15.4n3
AMD Radeon R9 M375
48
32
21.6
AMD Radeon R7 M275DX
49n2
22.65n2
15.2n2
15.85n2
AMD Radeon R7 512 Cores (Kaveri Desktop)
65
24
15.6
14
Intel Iris Pro Graphics 5200
53.3
24.2
16.3
13.6
NVIDIA GeForce GT 740M
61.4
23
14.5
12
NVIDIA GeForce 930M
57.3
50.6n2
19.5
NVIDIA GeForce 830M
50.6
21.2
14.9
12.5
AMD Radeon R7 M260X
40.8
18.2
12.8
11.2
AMD Radeon R7 M265
25
19.3
13
NVIDIA GeForce GT 640M
43.8
17.4
AMD Radeon R7 (Kaveri)
51.4
21.8
15.6
13.1
AMD Radeon R8 M365DX
30.2
19.5
14.5
12.8
AMD Radeon R7 M360
38.2
31.2
20.8
15.6
PosModel< PrevNext >Civilization: Beyond Earth
low med. high ultra
AMD Radeon R7 M260
29.3
21.7
15
Intel HD Graphics 530
57.7
22.3
13.1
10.9
AMD Radeon R6 (Carrizo)
32.4
19
11.3
8.8
AMD Radeon HD 8670M
22.2
15.9
AMD Radeon R6 M255DX
34.4
12.3
7.4
6.5
AMD Radeon HD 8650G
38.7
14.9
NVIDIA GeForce GT 630M
35.6
13
7.7
6.7
AMD Radeon R5 M255
62
26.9
16
15.6
NVIDIA GeForce 820M
36.3
14.3
Intel Iris Graphics 6100
43.4
23.2
15.9
12.6
NVIDIA GeForce GT 720M
44.8
15.4
9
AMD Radeon R5 M240
33.4
18.4
11.6
AMD Radeon R5 M230
17.5
12.6
7.8
Intel HD Graphics 6000
40.9
20.8
13.3
10.7
Intel Iris Graphics 5100
32.3n2
13.85n2
9.95n2
Intel HD Graphics 4600
35.65n2
13.35n2
8.65n2
7.4
Intel HD Graphics 5500
38.3n5
18.1n5
10.5n4
8
AMD Radeon HD 7660G
20
Intel HD Graphics 5000
15.5
12.2
Intel HD Graphics 4400
23.5
10
PosModel< PrevNext >Civilization: Beyond Earth
low med. high ultra
Intel HD Graphics 5300
21.2
8.8
5.8
4.5
Intel HD Graphics 4000
18.2
8.1
AMD Radeon R5 (Beema/Carrizo-L)
32.1
10.1
AMD Radeon R4 (Beema)
33.2
10.2
AMD Radeon R3 (Mullins/Beema)
27.6n2
9.5n2
Intel HD Graphics 4200
7.3
4.2
AMD Radeon HD 8350G
25.8
8.7
Intel HD Graphics (Braswell)
6
3.2
Intel HD Graphics (Cherry Trail)
16.2
6.1
Intel HD Graphics 3000
9.5
Intel HD Graphics (Bay Trail)
10.55n2
4.3
(-) * Smaller values are better. / n123 Number of benchmarks for this median value / * Approximate position

 

Legend
5Stutters – This game is very likely to stutter and have poor frame rates. Based on all known benchmarks using the specified graphical settings, average frame rates are expected to fall below 25fps
May Stutter – This graphics card has not been explicitly tested on this game. Based on interpolated information from surrounding graphics cards of similar performance levels, stutters and poor frame rates are expected.
30Fluent – Based on all known benchmarks using the specified graphical settings, this game should run at or above 25fps
40Fluent – Based on all known benchmarks using the specified graphical settings, this game should run at or above 35fps
60Fluent – Based on all known benchmarks using the specified graphical settings, this game should run at or above 58fps
May Run Fluently – This graphics card has not been explicitly tested on this game. Based on interpolated information from surrounding graphics cards of similar performance levels, fluent frame rates are expected.
?Uncertain – This graphics card experienced unexpected performance issues during testing for this game. A slower card may be able to achieve better and more consistent frame rates than this particular GPU running the same benchmark scene.
Uncertain – This graphics card has not been explicitly tested on this game and no reliable interpolation can be made based on the performances of surrounding cards of the same class or family.
The value in the fields displays the average frame rate of all values in the database. Move your cursor over the value to see individual results.
static version load dynamic
Loading Comments
Comment on this article
Please share our article, every link counts!
Florian Glaser, 2014-11-10 (Update: 2021-05-18)